Retratos y memorias de un pueblo en las selvas
118 ] Por el agujero de la memoria construyendo PAZ
Our final analysis examines the implications of stale performance chasing for investors. To an extent, some of the implications of stale performance chasing behavior are already well understood. Regardless of whether investors are reacting to contemporaneous or lagged performance signals, as shown by Berk and Green (2004) and Chen et al. (2004), as assets under management expand, economies of scale diminish resulting in a decline in performance with fund size. Lou (2012) similarly documents significant and temporary price impacts of flow-induced trading across mutual funds, which reverse in subsequent years. However, other implications may
exist beyond scale effects. Mutual fund fees are set in a competitive environment. Funds that wish to increase fees are likely to time the increase to special situations. For example, Bris, Gulen, Kadiyala, and Rau (2006) show that mutual funds choose to raise fees at points in time when they close their funds to new investors. Thus, beyond the understood performance implications, stale performance chasing may provide mutual funds with the opportunity to increase fees, in essence cashing in on the mechanical effects of HPRs.
To test this hypothesis, we relate fees to our return-chasing estimates. Following Khorana, Servaes, and Tufano (2009) who examine the determinants of mutual fund fees around the world, in these models we include as controls: 1) fund and family-level visibility proxies including size, age and the number of funds in the family, 2) fund change in market share and performance as funds and families with stronger performance and asset inflows are in a stronger competitive position to charge higher fees, 3) fund fees, 4) fund market share standard deviation as a proxy for fund value uncertainty and 5) the aggregate TNA of the funds in the same objective classification as a proxy for the level of competition between funds with a given management strategy. Fund and family performance is measured using Fama and French (1993) and Carhart (1997) 4-factor alphas, with family performance measured as the average alpha to all equity funds in the family. The results are reported in Table VI.
We use two return-chasing measures. The first is calculated as in equation (1) but using fund return in excess of the objective value-weighted mean return (Excess βn). The second is calculated by relating aggregate change in market share to all funds in objective j to the objective
j value-weighted mean return (Objective βn). These separate return-chasing measures allow us to differentiate between idiosyncratic and objective based return-chasing effects of fees.
Focusing first on the %Fee regressions in Panel A of Table VI, we find that funds with higher stale return-chasing in the prior period typically charge higher fees as a percentage of TNA. The magnitude of the effect is similar between fund and objective-level stale return signals. On average, a 1 standard deviation increase in return-chasing (Excess βn) across the three models is associated with a 10.5% increase in fees. A joint 1 standard deviation shift in Excess βn and
higher alphas, higher investor allocations (proxied by change in market share) and higher Morningstar ratings typically charge higher fees. Fund size and family size is typically related to lower fees, consistent with size being related to stronger governance (Guercio, Dan and Partch, 2003).
Although contractual fees are typically time invariant and may only be changed with shareholder consent, mutual funds routinely voluntarily waive fees during periods of poor performance in order to retain performance-sensitive investors (Christoffersen, 2001). To test this channel, we obtain fee waiver data from Morningstar, which reports the percentage of TNA waived from fees in a given year. We then relate the change in waived fees to the stale performance chasing estimates and control variables (models (4) to (6)). The stale performance chasing coefficients are all positive and significant, suggesting that funds reduce waived fees on the heels of stale performance chasing flow. The absolute sizes of the coefficient estimates across the model sets are highly similar, suggesting that adjustment to waivers explains the majority of the change in fees associated with stale performance chasing. Thus, our results suggest a channel by which managers are able to realize opportunities to return fees to prior levels without having to generate exceptional returns, i.e. by waiting for mechanical increases in HPRs realized as time passes.
The results in Panel A of Table VI suggest that mutual fees are higher for mutual funds held by more stale return-sensitive investors. We next explore a more direct linkage between HPRs and mutual fund price setting. In Panel B, we replicate the models from Panel A, relating the change in fund waivers to HPRs over the 1, 3 and 5 year horizon. The same set of control variables is included in Panel B, but we do not report the coefficients in the interest of brevity. Focusing first on the full model, we find an inverse relation between the change in waivers and HPRs at all three horizons, suggesting waivers decrease following increases in reported returns. As in Table III, we then partition the models by or (the end-return that drops from the three
and five year HPR, respectively). We expect that fee increases will be incrementally higher when the improvement in HPR results from relatively lower end-returns dropping from the horizon of analysis. This relation would be consistent with funds taking advantage of incremental fund demand resulting from end-return related improvements in HPRs which have been made more
visible to investors via advertising. Consistent with this thesis, we find statistically significant differences in the magnitude of waiver decreases associated with end-return related HPR increases. For both the three and five year HPR, fee increases are higher when investor demand is driven by stale as opposed to recent performance.
To summarize these results, we document a channel by which mutual funds exploit naïve or inattentive investors, benefiting from their inability to differentiate between the stale and current information components which jointly influence the change in HPR. Stale return chasing creates investor demand, in turn, creating opportunities for annulation of previously offered fee waivers. The resultant realized fees increase harms current investors who pay higher fees to hold the same fund. Quantifying harm to naïve investors who are enticed to purchase the fund via uninformative persuasion is more difficult, as the counterfactual performance of the investment they would have alternatively held cannot be observed. Unreported results suggest that the performance of funds affected by high stale return chasing is not significantly different from other funds, after controlling for fees, fund size and portfolio turnover (all underperform the market on average). However new investors also pay higher fees to hold the same fund. As previously discussed, all holders of the fund are also exposed to greater potential diseconomies of scale with its related drag on performance.
8. Conclusions
In this paper, we investigate if horizon effects in reported performance affect investor allocations to mutual funds. Specifically, we examine changes in reported HPR values in mutual fund advertisements. Changes in HPRs are equally influenced by the most recent return which enters and the end-return which drops from the horizon of the HPR calculation. Thus, as time passes, prior poor performance mechanically drops from the horizon of the HPR sample, potentially creating the misperception of improved fund performance.
Our analysis suggests that investors fail to recognize the effect of horizons on HPR calculations, allocating disproportionate wealth to funds when negative stale performance drops
from the horizon. Stale performance chasing is more pronounced in fund families that advertise HPR information and is more prevalent during periods of stress and uncertainty in financial markets. Fund families that advertise HPR information during these periods realize incrementally greater stale performance chasing behaviour. Either due to naivety or limited attention, investors misinterpret changes in HPRs, failing to distinguish between the relevant and stale information components of reported performance.
Finally, we show that mutual funds appear to time fee increases to coincide with periods of stale performance chasing, taking advantage of unsophisticated investors who do not appreciate the mechanical effect of time on HPRs. Our results suggest that the form in which mutual fund performance is advertised and resultant horizon effects are important influences of investor behavior.
References
Alti, Aydogan, Ron Kaniel, and Uzi Yoeli, 2006, Why do institutional investors chase return trends?, unpublished working paper, University of Texas.
Amir, Eli, 1993, The market valuation of accounting information: The case of postretirement benefits other than pensions, The Accounting Review 68, 703-724.
Barber, Brad, Terrance Odean, and Lu Zheng, 2000, The behavior of mutual fund investors, Unpublished working paper, University of California at Davis.
Barber, Brad, Terrance Odean, and Lu Zheng, 2005, Out of sight, out of mind: The effects of expenses on mutual fund flows, Journal of Business, 78, 2095-2120.
Berk, Jonathan, and Rick Green, 2004, Mutual fund flows and performance in rational markets,
Journal of Political Economy, 111, 1269-1295.
Bertrand, Marianne, Dean Karlan, Sendhil Mullainathan, Eldar Shafir, and Jonathan Zinman, 2010, What’s advertising content worth? Evidence from a consumer credit marketing field experiment, Quarterly Journal of Economics, 125, 263-306.
Blake, Christopher and Matthew Morey, 2000, Morningstar ratings and mutual fund performance,
Journal of Financial and Quantitative Analysis 35, 451-486.
Bloomfield, Robert J., 2002, The “incomplete revelation hypothesis” and financial reporting,
Accounting Horizons, 16, 233-243.
Bris, Arturo, Huseyin Gulen, Padma Kadiyala, and P. Raghavendra Rau, 2007, Good stewards, cheap talkers, or family men? The impact of mutual fund closures on fund managers, flows, fees, and performance,” Review of Financial Studies, 20, 953-982.
Carhart, Mark, 1997, On persistence in mutual fund performance, Journal of Finance, 52, 57-82 Chalmers, John, Aditya Kaul, and Blake Phillips, 2013, The wisdom of crowds: Mutual fund
investors’ aggregate asset allocation decisions, Journal of Banking and Finance, 37, 3318- 3333.
Chen, Joseph, Harrison Hong, Ming Huang, and Jeffrey D. Kubik, 2004, Does fund size erode mutual fund performance? The role of liquidity and organization, American Economic
Review, 94, 1276-1302.
Chevalier, Judith and Glenn Ellison, 1997, Risk taking by mutual funds as a response to incentives,
Journal of Political Economy, 105, 1167-1200.
Christoffersen, Susan, 2001, Why do money fund managers voluntarily waive their fees?, Journal
Cooper, Michael J., Huseyin Gulen, and P. Raghavendra Rau, 2006, Changing names with style: Mutual fund name changes and their effects on fund flows, Journal of Finance, 60, 2825- 2858
Cremers, Martijn and Antti Petajisto, 2009, How active is your fund manager? A new measure that predicts performance, Review of Financial Studies 22, 3329-3365.
Del Guercio, Diane and Paula Tkac, 2008, Star power: The effect of Morningstar ratings on mutual fund flow, Journal of Financial and Quantitative Analysis 43, 907-936.
Del Guercio, Diane, Larry Dan and Meagan Partch, 2003, Governance and boards of directors in closed-end investment companies, Journal of Financial Economics 69, 111-159.
Fama, Eugene F. and Kenneth R. French, 1993, Common risk factors in the returns of stocks and bonds, Journal of Financial Economics 33, 3-56.
Fiske, Susan, 1995, Social cognition, In: Tesser, A. (Ed.), Advanced Social Psychology. Mc-Graw- Hill, New York, pp. 149-194.
Fiske, Susan and Shelly Taylor, 1991, Social Cognition, 2nd Edition, McGraw-Hill, New York
Gallaher, Steven, Ron Kaniel, and Laura T. Starks, 2008, Advertising and mutual funds: From families to individual funds, working paper, University of Texas.
Goyal, Amit and Sunil Wahal, 2013, Is momentum an echo?, unpublished working paper, University of Arizona.
Hand, John, 1990, A test of the extended functional fixation hypothesis, The Accounting Review 65, 740-763.
Heston, Steven L., and Ronnie Sadka, 2008, Seasonality in the cross-section of stock returns,
Journal of Financial Economics 87, 418–445.
Heston, Steven L., and Ronnie Sadka, 2010, Seasonality in the cross section of stock returns: the international evidence, Journal of Financial and Quantitative Analysis 45, 1133–1160. Hirshleifer, David, and Siew H. Teoh, 2003, Limited attention, information disclosure, and
financial reporting, Journal of Accounting and Economics, 36, 337-386.
Huang, Jennifer, Kelsey Wei and Hong Yan, 2007, Participation costs and the sensitivity of fund flows to past performance, Journal of Finance 62, 1273-1311.
Investment Company Institute (ICI), 2005, How funds use 12b-1 fees, Fundamentals, Investment Company Institute Research Brief, Vol. 14, No. 2.
Ippolito, Richard, 1992, Consumer reaction to measures of poor quality: Evidence from the mutual fund industry, Journal of Law and Economics 35, 45-70.
Jegadeesh, Narasimhan, 1990, Evidence of predictable behavior of security returns, Journal of
Finance 45, 881–898.
Kahneman, Daniel and Amos Tversky, 1973, On the psychology of prediction, Psychological
Review 80, 237-251.
Karceski, Jason, 2002, Returns-chasing behavior, mutual funds, and beta's death, Journal of
Financial and Quantitative Analysis, 37, 559-594.
Khorana, Ajay, Henri Servaes, and Peter Tufano, 2009, Mutual fund fees around the world, Review
of Financial Studies 22, 1279-1310.
Kraussl, Roman and Ralph Sandelowsky, 2007, The predictive performance of Morningstar’s mutual fund ratings, University of Amsterdam working paper.
Lewis, Jeffrey and Drew Linzer, 2005, Estimating regression models in which the dependent variable is based on estimates, Political Analysis 13, 345-364.
Li, Feng, 2008, Annual report readability, current earnings, and earnings persistence, Journal of
Accounting and Economics 45, 221-247.
Lou, Dong, 2012, A flow-based explanation for return predictability, Review of Financial Studies, 25, 3457-3489.
Lynch, Anthony W., and David K. Musto, 2003, How investors interpret past fund returns, Journal
of Finance, 58, 2033-2058.
Mullainathan, Sendhil, and Andrei Shleifer, 2005, Persuasion in finance, unpublished working paper, Harvard University.
Mullainathan, Sendhil, Joshua Schwartzstein, and Andrei Shleifer, 2008, Coarse thinking and persuasion, Quarterly Journal of Economics, 123, 577-619.
Palmiter, Alan R., and Ahmed E. Taha, 2012, Mutual fund performance advertising: Inherently and materially misleading?, Georgia Law Review 46, 289-.
Pumlee, Marlene, 2003, The effect of information complexity on analysts’ use of that information,
The Accounting Review 78, 275-296.
Reuter, Jonathan, and Eric Zitzewitz, 2006, Do ads influence editors? advertising and bias in the financial media, Quarterly Journal of Economics, 121, 197-227.
Salganik, Galla, 2012, The “smart-money” effect: Retail versus institutional mutual funds, Unpublished working paper, Ben-Gurion University.
Sirri, Erik R. and Peter Tufano, 1998, Costly search and mutual fund flows, Journal of Finance, 53, 1589-1622.
Spiegel, Matthew, and Hong Zhang, 2013, Mutual fund risk and market share-adjusted fund flows,
Journal of Financial Economics 108, 506-528.
White, Halbert, 1980, A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity, Econometrica 48, 817-838.
Table I
Stale Performance Chasing Measurement
This table reports the return-chasing coefficient (βn) from the series of 72 separate pooled OLS regressions: Panel A and C: ∆mt = α + β1Returnt-1 + βnReturnt-n + εt, and
Panel B: Flowt = α + β1Returnt-1 + βnReturnt-n + εt, for all values of n from 2 to 73. Each regression generates a Returnt-1 coefficient and an additional unique lagged return
coefficient (Returnt-n). The Returnt-n coefficient values are reported below. The average Returnt-1 coefficient value from the 72 regressions in Panel A is 0.31 (significant at
the 1% level). Two t-statistics are reported. The first (Reg. t-stat) tests the null H0: βn =0. The second (Diff. t-stat) tests the null H0: βn = ∑ | |. T-test statistics are
clustered by fund and date (month-year). In Panels A and C, the dependent variable mt, is market share defined as mutual fund total net assets (TNA) divided by the total
TNA of all funds in the sample in month t. In Panel B, the dependent variable is the percentage growth in mutual fund total net assets (TNA), defined as: (TNAi,t –TNAi,t- 1×(1+Returni,t))/TNAi,t-1 in month t for fund i. The sample includes all domestic, actively managed mutual funds in the U.S. between 1992 and 2010. Return coefficients
coinciding with the end of commonly reported holding period returns (1, 3, and 5 years) appear in bold face. The output for all 72 regressions are reported in Panels A-C. Panel D tests a parsimonious version of Panel C. In the interest of brevity, in Panel E only the coefficients of interest are reported (the full output is available in the online appendix to the paper). In Panel C, the return coefficients are estimated simultaneously in one model. In Panel E, in addition to market share, an additional proxy for investor preferences used in model 6: N-SAR net sales calculated as the difference between share sales and redemptions reported directly by the fund in form N-SAR. Controls included in select models in Panel E include log fund TNA, log fund age measured from the date of fund inception to the start of the year of βn estimation, log
number of funds offered by the investment company (family) of the fund, log total family TNA, fund return standard deviation calculated over the prior year based on the fund’s monthly returns, Morningstar rating, the change in Morningstar rating, portfolio turnover measured as the percentage of average total net assets traded in year t, and expense ratio measured as the percentage of total net assets which investors pay for the fund’s operating expenses.
Panel A: Base Model with market share
Rt-2 Rt-3 Rt-4 Rt-5 Rt-6 Rt-7 Rt-8 Rt-9 Rt-10 Rt-11 Rt-12 Rt-13 Coef. 0.09 0.10 0.16 0.08 -0.26 0.26 0.07 0.17 0.06 0.11 -0.19 -0.33 Reg. t-stat 1.40 1.52 2.24 1.32 -3.14 3.19 1.01 2.37 0.86 1.67 -2.79 -4.11 Diff. t-stat -0.80 -0.73 0.25 -0.98 1.35 1.39 -1.16 0.32 -1.15 -0.58 0.66 2.30 N 425,781 423,464 421,101 418,218 415,466 413,988 411,269 408,523 405,527 402,662 400,511 397,586 Adjusted R2 24.85% 24.64% 24.41% 24.18% 23.94% 23.76% 23.58% 23.46% 23.34% 23.21% 23.04% 22.85% Rt-14 Rt-15 Rt-16 Rt-17 Rt-18 Rt-19 Rt-20 Rt-21 Rt-22 Rt-23 Rt-24 Rt-25 Coef. -0.29 0.10 0.08 0.19 -0.27 0.17 0.09 0.08 -0.22 -0.11 -0.18 -0.27 Reg. t-stat -3.91 1.44 1.22 2.62 -3.73 2.46 1.41 1.33 -3.02 -1.66 -2.51 -3.33 Diff. t-stat 1.94 -0.73 -1.03 0.57 1.73 0.35 -0.76 -0.98 1.07 -0.61 0.48 1.51 N 403,507 400,578 398,413 395,794 394,029 391,120 389,669 387,184 385,720 382,822 381,436 378,767 Adjusted R2 22.79% 22.67% 22.54% 22.38% 22.17% 21.97% 21.80% 21.65% 21.49% 21.31% 21.17% 20.99% Rt-26 Rt-27 Rt-28 Rt-29 Rt-30 Rt-31 Rt-32 Rt-33 Rt-34 Rt-35 Rt-36 Rt-37 Coef. 0.08 0.23 0.12 0.13 0.09 0.14 0.15 0.01 -0.01 -0.19 -0.25 -0.43 Reg. t-stat 1.20 3.10 1.74 1.86 1.39 1.96 2.03 0.33 -0.14 -2.55 -3.12 -4.30 Diff. t-stat -1.06 1.12 -0.29 -0.25 -0.88 -0.01 0.01 -4.26 -2.07 0.55 1.33 2.85 N 380,512 377,904 375,120 373,385 371,617 369,063 366,879 365,162 363,731 361,676 359,114 357,322 Adjusted R2 20.36% 20.16% 20.05% 19.91% 19.78% 19.66% 19.48% 19.29% 19.10% 18.98% 18.81% 18.65%
Rt-38 Rt-39 Rt-40 Rt-41 Rt-42 Rt-43 Rt-44 Rt-45 Rt-46 Rt-47 Rt-48 Rt-49 Coef. -0.19 -0.19 0.12 0.10 0.13 -0.18 -0.16 -0.27 0.13 0.15 0.13 -0.17 Reg. t-stat -2.95 -2.95 1.68 1.64 1.92 -2.52 -2.12 -3.54 1.92 2.06 1.87 -2.49 Diff. t-stat 0.73 0.74 -0.42 -0.63 -0.22 0.49 0.17 1.61 -0.17 0.02 -0.24 0.41 N 348,012 344,491 341,421 338,142 334,007 331,006 328,159 324,349 320,266 317,078 313,458 310,568 Adjusted R2 18.20% 18.02% 17.91% 17.78% 17.67% 17.52% 17.40% 17.28% 17.11% 17.00% 16.83% 16.72% Rt-50 Rt-51 Rt-52 Rt-53 Rt-54 Rt-55 Rt-56 Rt-57 Rt-58 Rt-59 Rt-60 Rt-61 Coef. 0.10 0.09 0.18 0.07 0.15 0.08 -0.12 -0.19 -0.13 0.19 0.14 -0.28 Reg. t-stat 1.56 1.34 2.53 1.01 2.10 1.11 -1.72 -2.77 -1.80 2.92 1.92 -3.74 Diff. t-stat -0.68 -0.85 0.53 -1.08 0.09 -1.01 -0.35 0.64 -0.27 0.69 -0.08 1.80 N 308,801 304,975 302,097 298,173 295,673 292,895 290,368 287,747 284,090 280,348 277,562 274,148 Adjusted R2 16.55% 16.45% 16.30% 16.19% 16.03% 15.91% 15.76% 15.66% 15.55% 15.41% 15.29% 15.21% Rt-62 Rt-63 Rt-64 Rt-65 Rt-66 Rt-67 Rt-68 Rt-69 Rt-70 Rt-71 Rt-72 Rt-73 Coef. 0.08 0.07 0.12 0.09 0.09 0.06 0.07 -0.15 -0.10 0.08 0.10 0.06 Reg. t-stat 1.26 0.99 1.71 1.33 1.40 0.91 1.07 -2.07 -1.63 1.28 1.50 0.87 Diff. t-stat -1.07 -1.21 -0.4 -0.86 -0.85 -1.16 -1.09 0.08 -0.67 -0.99 -0.73 -1.15 N 270,880 268,244 264,796 261,989 258,582 255,066 251,679 248,800 245,620 243,126 240,018 237,237 Adjusted R2 14.76% 14.62% 14.49% 14.35% 14.25% 14.16% 14.03% 13.95% 13.87% 13.75% 13.67% 13.58%
Panel B: Base model with flows
Rt-2 Rt-3 Rt-4 Rt-5 Rt-6 Rt-7 Rt-8 Rt-9 Rt-10 Rt-11 Rt-12 Rt-13 Coef. 0.11 0.09 0.14 0.09 -0.22 0.23 0.09 0.21 0.08 0.10 -0.20 -0.28 Reg. t-stat 1.53 1.23 2.08 1.29 -3.04 3.06 1.30 2.92 1.00 1.52 -2.67 -3.60 Diff. t-stat -0.47 -0.73 -0.05 -0.77 -1.06 -1.15 -0.78 -0.92 -0.8 -0.66 -0.75 -1.76 N 425,781 423,464 421,101 418,218 415,466 413,988 411,269 408,523 405,527 402,662 400,511 397,586 Adjusted R2 25.55% 25.32% 25.14% 24.93% 24.80% 24.64% 24.43% 24.28% 24.11% 23.91% 23.77% 23.61% Rt-14 Rt-15 Rt-16 Rt-17 Rt-18 Rt-19 Rt-20 Rt-21 Rt-22 Rt-23 Rt-24 Rt-25 Coef. -0.21 0.08 0.10 0.23 -0.23 0.14 0.11 0.11 -0.18 -0.09 -0.18 -0.27 Reg. t-stat -2.94 1.02 1.52 3.13 -3.17 2.09 1.53 1.54 -2.43 -1.33 -2.43 -3.59 Diff. t-stat -0.93 -0.81 -0.66 -1.18 -1.19 -0.05 -0.47 -0.47 -0.49 -0.79 -0.49 -1.68 N 403,507 400,578 398,413 395,794 394,029 391,120 389,669 387,184 385,720 382,822 381,436 378,767 Adjusted R2 22.98% 22.83% 22.64% 22.50% 22.36% 22.18% 22.03% 21.81% 21.67% 21.47% 21.30% 21.13% Rt-26 Rt-27 Rt-28 Rt-29 Rt-30 Rt-31 Rt-32 Rt-33 Rt-34 Rt-35 Rt-36 Rt-37 Coef. 0.09 0.19 0.12 0.15 0.08 0.12 0.15 0.01 -0.01 -0.21 -0.21 -0.38 Reg. t-stat 1.37 2.53 1.84 2.14 1.03 1.85 2.24 0.21 -0.34 -2.94 -3.00 -3.78 Diff. t-stat -0.82 -0.62 -0.36 -0.09 -0.82 -0.36 -0.1 -2.75 -4.51 -0.93 -0.95 2.35 N 380,512 377,904 375,120 373,385 371,617 369,063 366,879 365,162 363,731 361,676 359,114 357,322 Adjusted R2 23.29% 23.09% 22.90% 22.71% 22.52% 22.32% 22.19% 21.97% 21.78% 21.58% 21.40% 0.21
Rt-38 Rt-39 Rt-40 Rt-41 Rt-42 Rt-43 Rt-44 Rt-45 Rt-46 Rt-47 Rt-48 Rt-49 Coef. -0.22 -0.16 0.11 0.09 0.14 -0.20 -0.18 -0.23 0.11 0.18 0.11 -0.19 Reg. t-stat -3.04 -2.42 1.61 1.39 2.12 -2.73 -2.45 -3.58 1.63 2.45 1.70 -2.54 Diff. t-stat -1.06 -0.25 -0.49 -0.83 -0.05 -0.77 -0.5 -1.34 -0.5 -0.5 -0.52 -0.62 N 348,012 344,491 341,421 338,142 334,007 331,006 328,159 324,349 320,266 317,078 313,458 310,568 Adjusted R2 21.55% 21.42% 21.31% 21.19% 21.04% 20.87% 20.71% 20.51% 20.38% 20.25% 20.09% 19.96% Rt-50 Rt-51 Rt-52 Rt-53 Rt-54 Rt-55 Rt-56 Rt-57 Rt-58 Rt-59 Rt-60 Rt-61 Coef. 0.13 0.11 0.19 0.08 0.19 0.09 -0.13 -0.20 -0.13 0.18 0.12 -0.28 Reg. t-stat 1.98 1.72 2.59 1.04 2.67 1.41 -2.02 -2.90 -2.03 2.53 1.86 -3.77 Diff. t-stat -0.21 -0.53 -0.63 -0.83 -0.65 -0.84 -0.21 -0.82 -0.21 -0.51 -0.37 -1.84 N 308,801 304,975 302,097 298,173 295,673 292,895 290,368 287,747 284,090 280,348 277,562 274,148 Adjusted R2 21.39% 21.27% 21.12% 20.94% 20.77% 20.62% 20.41% 20.27% 20.13% 19.97% 19.79% 19.67% Rt-62 Rt-63 Rt-64 Rt-65 Rt-66 Rt-67 Rt-68 Rt-69 Rt-70 Rt-71 Rt-72 Rt-73 Coef. 0.08 0.07 0.11 0.09 0.10 0.08 0.08 -0.16 -0.10 0.06 0.10 0.07 Reg. t-stat 1.05 0.94 1.79 1.21 1.44 1.16 1.17 -2.28 -1.43 0.91 1.45 0.96 Diff. t-stat -0.78 -1.04 -0.54 -0.74 -0.66 -0.85 -0.83 -0.22 0.67 -1.3 -0.64 -1.01 N 270,880 268,244 264,796 261,989 258,582 255,066 251,679 248,800 245,620 243,126 240,018 237,237 Adjusted R2 19.70% 19.59% 19.42% 19.28% 19.09% 18.98% 18.83% 18.67% 18.51% 18.38% 18.27% 18.11%
Panel C: Simultaneous Estimation Model (N=237,237, Adj. R2=27.90)
Rt-2 Rt-3 Rt-4 Rt-5 Rt-6 Rt-7 Rt-8 Rt-9 Rt-10 Rt-11 Rt-12 Rt-13 Coef. 0.12 0.11 0.15 -0.13 -0.40 -0.33 0.09 0.14 -0.05 0.10 -0.31 -0.42 Reg. t-stat 1.70 1.51 2.15 -1.99 -4.22 -4.18 1.28 2.05 -0.73 1.39 -3.88 -4.38 Diff. t-stat -0.34 -0.43 0.19 -0.16 2.75 2.40 -0.71 -0.05 -1.26 -0.62 2.10 2.91 Rt-14 Rt-15 Rt-16 Rt-17 Rt-18 Rt-19 Rt-20 Rt-21 Rt-22 Rt-23 Rt-24 Rt-25 Coef. 0.20 0.07 0.07 -0.13 -0.15 -0.17 0.10 0.12 0.13 0.05 -0.14 -0.21 Reg. t-stat 2.88 0.99 1.00 -1.98 -2.08 -2.16 1.41 1.73 1.98 0.72 -2.02 -2.89 Diff. t-stat 0.86 -1.13 -1.13 -0.21 0.06 0.31 -0.60 -0.33 -0.18 -1.43 -0.09 0.92 Rt-26 Rt-27 Rt-28 Rt-29 Rt-30 Rt-31 Rt-32 Rt-33 Rt-34 Rt-35 Rt-36 Rt-37 Coef. -0.16 0.18 0.14 -0.15 -0.11 0.13 0.11 -0.01 -0.01 0.17 -0.18 -0.29 Reg. t-stat -2.15 2.57 2.03 -2.13 -1.50 1.86 1.50 -0.08 -0.10 2.20 -2.54 -3.51 Diff. t-stat 0.25 0.58 -0.04 0.18 -0.44 -0.23 -0.47 -0.90 -1.02 0.33 0.54 1.78 Rt-38 Rt-39 Rt-40 Rt-41 Rt-42 Rt-43 Rt-44 Rt-45 Rt-46 Rt-47 Rt-48 Rt-49 Coef. 0.21 0.15 0.06 0.12 0.17 -0.18 -0.13 0.18 0.14 -0.12 -0.12 -0.10 Reg. t-stat 2.99 2.12 0.84 1.64 2.33 -2.62 -1.87 2.85 2.04 -1.81 -1.60 -1.33 Diff. t-stat 1.00 0.09 -1.01 -0.34 0.38 0.59 -0.23 0.65 -0.01 -0.30 -0.35 -0.62
Rt-50 Rt-51 Rt-52 Rt-53 Rt-54 Rt-55 Rt-56 Rt-57 Rt-58 Rt-59 Rt-60 Rt-61 Coef. -0.09 -0.12 0.18 0.07 0.24 0.10 0.20 0.20 -0.13 0.21 -0.11 -0.30 Reg. t-stat -1.16 -1.82 2.43 0.92 3.15 1.39 2.85 2.85 -1.93 2.99 -1.50 -3.66 Diff. t-stat -0.72 -0.30 0.49 -1.07 1.33 -0.63 0.83 0.79 -0.23 0.98 -0.45 1.95 Rt-62 Rt-63 Rt-64 Rt-65 Rt-66 Rt-67 Rt-68 Rt-69 Rt-70 Rt-71 Rt-72 Rt-73 Coef. 0.09 0.06 0.12 0.08 0.11 0.07 0.07 -0.12 -0.10 0.06 0.10 0.06 Reg. t-stat 1.29 0.85 1.60 1.10 1.53 1.10 1.08 -1.67 -1.48 0.90 1.43 0.79 Diff. t-stat -0.66 -1.02 -0.37 -0.84 -0.39 -0.97 -0.96 -0.34 -0.59 -1.08 -0.59 -1.23
Panel D: Base model with flows
Model (1) (2) Rt 0.45*** (4.17) Rt-1 0.39*** 0.34*** (4.43) (3.54) Rt-1,t-12 0.06** 0.06** (2.16) (2.23) Rt-13 -0.30*** -0.32*** (-3.16) (-3.71) Rt-15,t-36 0.01** 0.01** (2.27) (2.37) Rt-37 -0.37*** -0.40*** (-4.64) (-3.78) R t-39,t-60 0.01 0.02** (1.85) (2.37) Rt-61 -0.26*** -0.29*** (-3.27) (-4.01) N 237,237 237,237 Adjusted R2 36.21% 38.17%
Coefficients are reported with difference t-statistics (as defined in Table 1 of the main paper) reported in brackets. Coefficients significant at the 1% and 5% levels are marked with *** and ** respectively.
Panel E: Robustness Tests and Extensions
Model Dep. Var. Clustered S.E. Fixed Effects Controls Rt-13 Rt-37 Rt-61
(1) ∆ Market Share Fund, Date No No -0.33 -0.43 -0.28 (-4.11) (-2.85) (-3.74) (2) ∆ Market Share Fund, Time, Family No No -0.33 -0.43 -0.28
(-3.62) (-4.44) (-3.45) (3) ∆ Market Share Fund, Date No Yes -0.34 -0.36 -0.23
(-3.79) (-4.31) (-2.96) (4) ∆ Market Share Fund, Date Fund Objective No -0.26 -0.35 -0.30
(-3.72) (-3.44) (-3.20) (5) ∆ Market Share Fund, Date Fund Objective Yes -0.33 -0.34 -0.35
(-3.58) (-4.11) (-4.63) (6) N-SAR Net Sales Fund, Date No No -0.34 -0.39 -0.41
(-3.61) (-3.68) (-3.75)
Interpretation: This table demonstrates that investors allocate disproportionate capital to funds with negative returns 13, 37, and 61 months prior. These returns coincide with the end of commonly reported holding periods for mutual funds, suggesting that investors
are unable to differentiate between improvements in HPRs which result from high recent returns (Returnt-1) relative to improvements
Table II
Advertising Expenditure Dataset Summary Statistics
This table summarizes investment company advertising expenditures in print media in the U.S. by year (thousand USD). Expenditures are segmented between magazine and newspaper advertisements. Magazine advertisement expenditures are subsequently segmented by whether the advertisement promotes mutual funds or promotes another product or service provided by the mutual fund company. Expenditures focused on mutual fund advertising are subsequently segmented by: 1) expenditures that promote a specific fund relative to broad family based promotion, 2) expenditures which promote holding period returns (HPR) by horizon, 3) expenditures which promote rankings provided by Morningstar or another analyst by horizon and 4) analyst rankings not by horizon. Greater detail regarding the establishment of the partitions is provided in the Appendix. 2005 2006 2007 2008 2009 2010 Total % of Total All Advertising Newspaper 61,427 58,794 89,764 96,690 103,630 126,881 537,186 34 Magazine 148,643 169,341 184,043 213,006 147,315 182,969 1,045,317 66 Total 210,070 228,135 273,807 309,696 250,945 309,850 1,582,503 Magazine Advertising
Not Mutual Fund Focused 84,061 96,872 110,181 149,840 123,985 147,823 712,762 68
Mutual Fund Focused 64,582 72,469 73,862 63,166 23,330 35,146 332,555 32
Mutual Fund Focused Magazine Advertising
Fund Specific 49,869 52,658 54,202 39,008 15,517 18,212 229,466 69
Family 14,713 19,811 19,660 24,158 7,813 16,934 103,089 31
Fund Specific Magazine Advertising
Advertise HPR by Horizon 28,535 23,605 20,857 13,666 1,615 1,675 89,953 39
Advertise Rank by Horizon 6,363 2,634 2,003 3,818 2,144 938 17,900 8
Advertise Analyst Rating 5,201 22,321 24,458 12,632 9,234 14,082 87,928 38
Other Advertising 9,770 4,098 6,884 8,892 2,524 1,517 33,685 15
Interpretation: Investment company advertising expenditure promoting mutual funds is cyclical. It is higher during periods of strong market performance.
Table III
Determinants of Holding Period Return Advertising Expenditures
In this table, the dependent variable is the change in spending on fund specific advertising that features holding period return data or analyst rating in year T standardized by fund size. The key independent variables are holding period returns (HPR) calculated using monthly returns over the prior 1, 3, and 5 year horizons (for example, refers to the 3 year HPR ending in year T-1). The models are partitioned by the annual return that drops from the 3 and 5 year HPR horizon (i.e. the 1 year HPR